Clean up ppp_stdlib.py

I will need to implement polymorphism later to allow for functions like `len` to work. Len currently relies on saying that its argument is either a list or a string, as strings are not a subtype of lists.

ppp_ast.py

@@ -1,6 +1,6 @@
 from abc import ABC, abstractmethod
 from dataclasses import dataclass
-from typing import Dict, List, Optional, Tuple, Union
+from typing import Dict, List, Optional, Tuple
 
 ### Types ###
 
@@ -15,13 +15,6 @@ class TupleTypeExpr(TypeExpression):
 	def represent(self) -> str:
 		assert False, ("Unimplemented")
 
-@dataclass
-class UnionTypeExpr(TypeExpression):
-	types: List[TypeExpression]
-
-	def represent(self) -> str:
-		assert False, ("Unimplemented")
-
 @dataclass
 class ListTypeExpr(TypeExpression):
 	type: TypeExpression
@@ -144,6 +137,7 @@ class ArrayAccess(Expression):
 
 @dataclass
 class Array(Expression):
+	element_type: TypeExpression
 	array: List[Expression]
 
 	def represent(self) -> str:
@@ -200,6 +194,7 @@ class StructInstantiation(Expression):
 
 @dataclass
 class LoopComprehension(Expression):
+	element_type: TypeExpression
 	body: Expression
 	variable: str # TODO: Pattern matching
 	array: Expression
@@ -209,16 +204,6 @@ class LoopComprehension(Expression):
 
 	def precedence(self) -> int: return 13
 
-@dataclass
-class Return(Expression):
-	expression: Expression
-
-	def represent(self) -> str:
-		# TODO: This will have to be improved
-		return "return "+self.wrap(self.expression)
-
-	def precedence(self) -> int: return 0
-
 @dataclass
 class Lambda(Expression):
 	parameters: List[TypeDeclaration]
@@ -491,7 +476,6 @@ class DoWhileStatement(Statement):
 	body: Statement
 	condition: Optional[Expression]
 
-# TODO: Maybe do something similar to return with these two?
 @dataclass
 class BreakStatement(Statement):
 	pass
@@ -500,6 +484,10 @@ class BreakStatement(Statement):
 class ContinueStatement(Statement):
 	pass
 
+@dataclass
+class ReturnStatement(Statement):
+	expression: Expression
+
 @dataclass
 class MatchStatement(Statement):
 	value: Expression
@@ -518,7 +506,7 @@ class ForLoop(Statement):
 
 @dataclass
 class Import(Statement):
-	file: Expression
+	file: str
 
 @dataclass
 class TypeDefinition(Statement):

ppp_interpreter.py

@@ -1,13 +1,13 @@
 from dataclasses import dataclass
-from typing import Dict, List as List_, Optional, Tuple, Union
+from typing import Dict, List as List_, Optional, Tuple
 
 from ppp_ast import *
 from ppp_lexer import Lexer
-from ppp_object import Bool, EnumValue, Function, Int, Object, Str, Struct, Tuple as TupleObject, List as ListObject, Return as ReturnObject, TypeObject, Void
+from ppp_object import Bool, EnumValue, Function, Int, Object, Str, Struct, Tuple as TupleObject, List as ListObject, TypeObject, Void
 from ppp_parser import is_valid_target, parse_statement
 from ppp_tokens import EofToken
 from ppp_stdlib import variables
-from ppp_types import EnumType, FunctionType, GenericType, Int as IntType, ListType, ReturnType, Str as StrType, StructType, TupleType, Type, TypeType, UnionType, VariableType, Void as VoidType
+from ppp_types import EnumType, FunctionType, GenericType, Int as IntType, ListType, Str as StrType, StructType, TupleType, Type, TypeType, VariableType, Void as VoidType
 
 @dataclass
 class Declared:
@@ -31,7 +31,7 @@ class Constant:
 	def from_obj(obj: Object) -> 'Declared':
 		return Declared(obj.get_type(), obj)
 
-VariableState = Union[Declared, Undeclared, Constant]
+VariableState = Declared | Undeclared | Constant
 
 Module = Dict[str, VariableState]
 
@@ -218,9 +218,6 @@ def calculate_expression(expression: Expression, program: ProgramState) -> Objec
 					case Int(num): return Str(left_value.str % num)
 					case _: assert False, ("Unimplemented", right_value)
 			assert False, ("Unimplemented", lhs, rhs)
-		case Return(expression):
-			value = calculate_expression(expression, program)
-			return ReturnObject(ReturnType(value.get_type()), value)
 		case StructInstantiation(struct_, arguments_):
 			struct = calculate_expression(struct_, program)
 			assert isinstance(struct, TypeObject)
@@ -291,38 +288,26 @@ def calculate_expression(expression: Expression, program: ProgramState) -> Objec
 			assert False, ("Unimplemented", expression_)
 		case UnaryMinus (expression_):
 			assert False, ("Unimplemented", expression_)
-		case Array(array_):
-			if len(array_) == 0:
-				return ListObject(ListType(VariableType("")), [])
-			elements_type: Optional[Type] = None
+		case Array(element_type_, array_):
+			element_type = calculate_type_expression(element_type_, program)
 			array_elements_: List_[Object] = []
 			for element_ in array_:
 				element = calculate_expression(element_, program)
-				if elements_type:
-					assert element.get_type().is_subtype_of(elements_type), (element, elements_type)
-				else:
-					elements_type = element.get_type()
+				assert element.get_type().is_subtype_of(element_type), (element, element_type)
 				array_elements_.append(element)
-			assert elements_type
-			return ListObject(ListType(elements_type), array_elements_)
-		case LoopComprehension(body_, variable, array_):
+			return ListObject(ListType(element_type), array_elements_)
+		case LoopComprehension(element_type_, body_, variable, array_):
+			element_type = calculate_type_expression(element_type_, program)
 			array = calculate_expression(array_, program)
 			assert array.get_type().is_indexable()
 			if isinstance(array, ListObject):
 				elements: List_[Object] = []
-				elements_type = None
 				for element in array.list:
 					program.push_context({variable: Declared.from_obj(element)})
 					elements.append(calculate_expression(body_, program))
 					program.pop_context()
-					if elements_type:
-						assert elements[-1].get_type().is_subtype_of(elements_type)
-					else:
-						elements_type = elements[-1].get_type()
-
-				if not elements: return ListObject(ListType(VariableType("")), [])
-				assert elements_type
-				return ListObject(ListType(elements_type), elements)
+					assert elements[-1].get_type().is_subtype_of(element_type)
+				return ListObject(ListType(element_type), elements)
 			else:
 				assert False, ("Unimplemented", array)
 		case _:
@@ -340,8 +325,6 @@ def calculate_type_expression(expression: TypeExpression, program: ProgramState,
 			return ListType(calculate_type_expression(type_, program, must_resolve))
 		case TupleTypeExpr(types_):
 			return TupleType([calculate_type_expression(type, program, must_resolve) for type in types_])
-		case UnionTypeExpr(types_):
-			return UnionType([calculate_type_expression(type, program, must_resolve) for type in types_])
 		case FunctionTypeExpr(arguments_, return_type_):
 			return FunctionType([calculate_type_expression(argument, program, must_resolve) for argument in arguments_], calculate_type_expression(return_type_, program, must_resolve))
 		case TypeSpecification(type_, types_):
@@ -413,14 +396,13 @@ class BreakResult:
 class NothingResult:
 	pass
 
-StatementsResult = Union[ReturnResult, ContinueResult, BreakResult, NothingResult]
+StatementsResult = ReturnResult | ContinueResult | BreakResult | NothingResult
 
 def interpret_statements(statements: List_[Statement], program: ProgramState) -> StatementsResult:
 	for statement in statements:
 		match statement:
 			case ExpressionStatement(expression):
-				value = calculate_expression(expression, program)
-				if isinstance(value, ReturnObject): return ReturnResult(value.value)
+				calculate_expression(expression, program)
 			case Assignment(lhs, rhs, type_):
 				assert is_valid_target(lhs)
 				match lhs:
@@ -554,14 +536,14 @@ def interpret_statements(statements: List_[Statement], program: ProgramState) ->
 							case _: assert False, ("Unimplemented", return_value)
 			case ContinueStatement(): return ContinueResult()
 			case BreakStatement(): return BreakResult()
-			case Import(file_):
+			case ReturnStatement(expression=expression):
+				return ReturnResult(calculate_expression(expression, program))
+			case Import(file):
 				# TODO: Maybe an inclusion system within a preprocessor maybe
-				file = calculate_expression(file_, program)
-				assert isinstance(file, Str), "Only strings are valid file paths!"
-				module = interpret_file(file.str, program.modules) if file.str not in program.modules else program.modules[file.str]
+				module = interpret_file(file, program.modules) if file not in program.modules else program.modules[file]
 				program.contexts[0] |= module
-				if file.str not in program.modules:
-					program.modules[file.str] = module
+				if file not in program.modules:
+					program.modules[file] = module
 			case TypeDefinition(name, expression_):
 				program.declare_and_assign_variable(name, TypeObject(calculate_type_expression(expression_, program)))
 			case DeferStatement(statement=statement):
@@ -582,7 +564,7 @@ def interpret_file(file_path: str, modules: Dict[str, Module]) -> Module:
 	assert len(program.contexts) == 2
 	match return_value:
 		case NothingResult(): pass
-		case ReturnObject(_): assert False, "Cannot return from outside a function!"
+		case ReturnResult(_): assert False, "Cannot return from outside a function!"
 		case ContinueResult(): assert False, "Cannot continue from outside a loop!"
 		case BreakResult(): assert False, "Cannot break from outside a loop!"
 		case _: assert False, ("Unimplemented", return_value)

ppp_lexer.py

@@ -1,31 +1,32 @@
 
 from typing import Optional
 
-from ppp_tokens import EofToken, IdentifierToken, Keyword, KeywordToken, NumberToken, StringToken, Symbol, SymbolToken, Token, TokenContents
+from ppp_tokens import EofToken, IdentifierToken, Keyword, KeywordToken, NumberToken, StringToken, Symbol, SymbolToken, Token, TokenContents, Location
 
 class Lexer:
-	def __init__(self, source: str) -> None:
+	def __init__(self, source: str, filename: str) -> None:
 		self._source = source
 		self._location = 0
 		self._line = 1
 		self._col = 0
+		self._filename = filename
 		self._peeked_token: Optional[Token] = None
-		self._current: str = ""
+
+	def _loc(self) -> Location:
+		return Location(self._filename, self._line, self._col)
+
+	def _token(self, loc: Location, value: str, contents: TokenContents) -> Token:
+		return Token(loc, value, contents)
 
 	@classmethod
 	def from_file(cls, path: str) -> 'Lexer':
 		with open(path) as f:
-			return cls(f.read())
+			return cls(f.read(), path)
 
-	def _advance(self) -> str:
+	def _advance(self):
 		assert self._location < len(self._source)
-		self._line, self._col = (self._line + 1, 0) if self._current == '\n' else (self._line, self._col + 1)
+		self._line, self._col = (self._line + 1, 0) if self._source[self._location] == '\n' else (self._line, self._col + 1)
 		self._location += 1
-		self._current = self._source[self._location] if self._location < len(self._source) else ''
-		return self._current
-
-	# def _peek(self) -> str:
-	# 	assert self._location < len(self._source)-1
 
 	def next_token(self) -> Token:
 		if self._peeked_token is not None:
@@ -34,71 +35,84 @@ class Lexer:
 
 		while self._location < len(self._source) and self._source[self._location] in ' \t\n': self._advance()
 
-		if self._location >= len(self._source): return Token(self._line, self._col, '\0', EofToken())
+		if self._location >= len(self._source): return self._token(self._loc(), '\0', EofToken())
 
 		match self._source[self._location]:
 			case c if c.isdigit():
 				start_location = self._location
-				while self._location < len(self._source) and self._source[self._location].isdigit(): self._location += 1
+				loc = self._loc()
+				while self._location < len(self._source) and self._source[self._location].isdigit(): self._advance()
 				number = int(self._source[start_location:self._location])
-				return Token(self._line, self._col, self._source[start_location:self._location], NumberToken(number))
+				return self._token(loc, self._source[start_location:self._location], NumberToken(number))
 			case c if c.isalpha() or c == "_":
 				start_location = self._location
-				while self._location < len(self._source) and (self._source[self._location].isalpha() or self._source[self._location] in '_'): self._location += 1
+				loc = self._loc()
+				while self._location < len(self._source) and (self._source[self._location].isalpha() or self._source[self._location] in '_'): self._advance()
 				word = self._source[start_location:self._location]
 				try:
 					keyword = Keyword(word)
-					return Token(self._line, self._col, word, KeywordToken(keyword))
+					return self._token(loc, word, KeywordToken(keyword))
 				except ValueError:
 					try:
 						symbol = Symbol(word)
-						return Token(self._line, self._col, word, SymbolToken(symbol))
+						return self._token(loc, word, SymbolToken(symbol))
 					except ValueError:
-						return Token(self._line, self._col, word, IdentifierToken(word))
+						return self._token(loc, word, IdentifierToken(word))
 			case '"':
-				# TODO: Escaping
-				self._location += 1
+				# TODO: Proper escaping
+				self._advance()
 				start_location = self._location
+				loc = self._loc()
 				escaping = False
 				while self._location < len(self._source) and (self._source[self._location] != '"' or escaping):
 					escaping = self._source[self._location] == '\\' if not escaping else False
-					self._location += 1
+					self._advance()
 				string = self._source[start_location:self._location].encode('utf-8').decode('unicode_escape')
-				self._location += 1
-				return Token(self._line, self._col, self._source[start_location-1:self._location], StringToken(string))
+				self._advance()
+				return self._token(loc, self._source[start_location-1:self._location], StringToken(string))
 			# TODO: Make a proper Trie for this.
 			case '|' if self._location < len(self._source)-1 and self._source[self._location+1] == '|':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.Dpipe))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.Dpipe))
 			case '&' if self._location < len(self._source)-1 and self._source[self._location+1] == '&':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.Dampersand))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.Dampersand))
 			case '*' if self._location < len(self._source)-1 and self._source[self._location+1] == '*':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.Dasterisk))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.Dasterisk))
 			case '-' if self._location < len(self._source)-1 and self._source[self._location+1] == '>':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.Arrow))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.Arrow))
 			case '>' if self._location < len(self._source)-1 and self._source[self._location+1] == '=':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.GreaterEqual))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.GreaterEqual))
 			case '<' if self._location < len(self._source)-1 and self._source[self._location+1] == '=':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.LesserEqual))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.LesserEqual))
 			case '=' if self._location < len(self._source)-1 and self._source[self._location+1] == '=':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.Dequal))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.Dequal))
 			case '=' if self._location < len(self._source)-1 and self._source[self._location+1] == '>':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.EqualArrow))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.EqualArrow))
 			case '!' if self._location < len(self._source)-1 and self._source[self._location+1] == '=':
-				self._location += 2
-				return Token(self._line, self._col, self._source[self._location-2:self._location], SymbolToken(Symbol.NotEqual))
+				loc = self._loc()
+				self._advance(); self._advance()
+				return self._token(loc, self._source[self._location-2:self._location], SymbolToken(Symbol.NotEqual))
 			case c if c in Symbol._value2member_map_:
-				self._location += 1
-				return Token(self._line, self._col, self._source[self._location-1], SymbolToken(Symbol(c)))
+				loc = self._loc()
+				self._advance()
+				return self._token(loc, self._source[self._location-1], SymbolToken(Symbol(c)))
 			case _:
-				assert False, ("Unimplemented", c, self._location)
+				raise SyntaxError(f"{self._loc()}: Unknown character: '{c}'")
 		assert False, "Unreachable"
 
 	def peek_token(self) -> Token:
@@ -108,12 +122,12 @@ class Lexer:
 
 	def assert_tokenkind(self, kind: type) -> Token:
 		token = self.next_token()
-		assert isinstance(token.contents, kind), (f"Expected {kind} but got {token.contents}!", self.next_token(), self.next_token(), self.next_token())
+		if not isinstance(token.contents, kind): raise SyntaxError(f"{token.loc}: Expected {kind} but got {token.contents}!")
 		return token
 
 	def assert_token(self, expected: TokenContents) -> Token:
 		token = self.next_token()
-		assert token.contents == expected, (f"Expected {expected} but got {token.contents}!", self.next_token(), self.next_token())
+		if token.contents != expected: raise SyntaxError(f"{token.loc}: Expected {expected} but got {token.contents}!")
 		return token
 
 	def check_token(self, expected: TokenContents) -> bool:

ppp_object.py

@@ -1,11 +1,9 @@
-# This file exists because I wanted to keep ppp_stdlib.py and ppp_interpreter.py seperate but they both rely on this one class.
-
 from abc import ABC, abstractmethod
 from dataclasses import dataclass
-from typing import Callable, Dict, List as List_, Tuple as Tuple_, Union as Union_
+from typing import Callable, Dict, List as List_, Tuple as Tuple_
 
 from ppp_ast import Statement
-from ppp_types import ArrayType, EnumType, FunctionType, ListType, ReturnType, StructType, TupleType, Type, Int as IntType, Str as StrType, Bool as BoolType, Void as VoidType, TypeType
+from ppp_types import ArrayType, EnumType, FunctionType, ListType, StructType, TupleType, Type, Int as IntType, Str as StrType, Bool as BoolType, Void as VoidType, TypeType
 
 class Object(ABC):
 	@abstractmethod
@@ -68,13 +66,6 @@ class Function(Object):
 
 	def get_type(self) -> Type: return self.type
 
-@dataclass
-class Return(Object):
-	type: ReturnType
-	value: Object
-
-	def get_type(self) -> Type: return self.type
-
 @dataclass
 class EnumValue(Object):
 	type: EnumType

ppp_parser.py

@@ -23,21 +23,10 @@ def parse_type_primary(lexer: Lexer) -> TypeExpression:
 	if lexer.take_token(SymbolToken(Symbol.Open)):
 		if lexer.take_token(SymbolToken(Symbol.Close)): return TupleTypeExpr([])
 
-		def parse_union(lexer: Lexer) -> TypeExpression:
-			union_types: List[TypeExpression] = [parse_type(lexer)]
-			while lexer.take_token(SymbolToken(Symbol.Pipe)):
-				union_types.append(parse_type(lexer))
-			if len(union_types) == 1:
-				return union_types[0]
-			return UnionTypeExpr(union_types)
-
-		types: List[TypeExpression] = [parse_union(lexer)]
+		types: List[TypeExpression] = [parse_type(lexer)]
 		while lexer.take_token(SymbolToken(Symbol.Comma)):
-			types.append(parse_union(lexer))
+			types.append(parse_type(lexer))
 		lexer.assert_token(SymbolToken(Symbol.Close))
-		if len(types) == 1 and isinstance(types[0], UnionTypeExpr):
-			base_type = types[0]
-		else:
 		base_type = TupleTypeExpr(types)
 	elif lexer.take_token(SymbolToken(Symbol.OpenSquare)):
 		type = parse_type(lexer)
@@ -119,21 +108,25 @@ def parse_primary(lexer: Lexer) -> Expression:
 			else:
 				base_expression = elements[0]
 	elif lexer.take_token(SymbolToken(Symbol.OpenSquare)):
+		lexer.assert_token(SymbolToken(Symbol.Colon))
+		element_type = parse_type(lexer)
+		
 		if lexer.take_token(SymbolToken(Symbol.CloseSquare)):
-			base_expression = Array([])
+			base_expression = Array(element_type, [])
 		else:
+			lexer.assert_token(SymbolToken(Symbol.Comma))
 			expressions: List[Expression] = [parse_expression(lexer)]
 			if lexer.take_token(KeywordToken(Keyword.For)):
 				variable = parse_identifier(lexer) # TODO: Pattern matching
 				lexer.assert_token(KeywordToken(Keyword.In))
 				expression = parse_expression(lexer)
 				lexer.assert_token(SymbolToken(Symbol.CloseSquare))
-				base_expression = LoopComprehension(expressions[0], variable, expression)
+				base_expression = LoopComprehension(element_type, expressions[0], variable, expression)
 			else:
 				while lexer.take_token(SymbolToken(Symbol.Comma)):
 					expressions.append(parse_expression(lexer))
 				lexer.assert_token(SymbolToken(Symbol.CloseSquare))
-				base_expression = Array(expressions)
+				base_expression = Array(element_type, expressions)
 	elif lexer.check_tokenkind(StringToken):
 		base_expression = String(parse_string(lexer))
 	elif lexer.check_tokenkind(NumberToken):
@@ -141,26 +134,17 @@ def parse_primary(lexer: Lexer) -> Expression:
 	else:
 		base_expression = Variable(parse_identifier(lexer))
 
-	while (token := lexer.take_tokens(SymbolToken(Symbol.Open), SymbolToken(Symbol.OpenSquare), SymbolToken(Symbol.Dot), SymbolToken(Symbol.OpenCurly))):
+	while (token := lexer.take_tokens(SymbolToken(Symbol.Open), SymbolToken(Symbol.OpenSquare), SymbolToken(Symbol.Dot))):
 		match token.contents:
 			case SymbolToken(symbol):
 				match symbol:
 					case Symbol.Dot:
-						field = parse_identifier(lexer)
+						next_token = lexer.next_token()
+						match next_token.contents:
+							case IdentifierToken(identifier=field):
 								base_expression = FieldAccess(base_expression, field)
-					case Symbol.Open:
-						if lexer.take_token(SymbolToken(Symbol.Close)):
-							base_expression = FunctionCall(base_expression, [])
-						else:
-							arguments: List[Expression] = [parse_expression(lexer)]
-							while lexer.take_token(SymbolToken(Symbol.Comma)):
-								arguments.append(parse_expression(lexer))
-							lexer.assert_token(SymbolToken(Symbol.Close))
-							base_expression = FunctionCall(base_expression, arguments)
-					case Symbol.OpenSquare:
-						index = parse_expression(lexer)
-						lexer.assert_token(SymbolToken(Symbol.CloseSquare))
-						base_expression = ArrayAccess(base_expression, index)
+							case SymbolToken(symbol=symbol):
+								match symbol:
 									case Symbol.OpenCurly:
 										if lexer.take_token(SymbolToken(Symbol.CloseCurly)):
 											base_expression = StructInstantiation(base_expression, [])
@@ -174,6 +158,23 @@ def parse_primary(lexer: Lexer) -> Expression:
 											while lexer.take_token(SymbolToken(Symbol.Comma)): struct_arguments.append(parse_argument())
 											lexer.assert_token(SymbolToken(Symbol.CloseCurly))
 											base_expression = StructInstantiation(base_expression, struct_arguments)
+									case _:
+										raise SyntaxError(f"{next_token.loc}: Unexpected symbol: {repr(str(symbol))}")
+							case _:
+								raise SyntaxError(f"{next_token.loc}: Unexpected: {next_token.contents}")
+					case Symbol.Open:
+						if lexer.take_token(SymbolToken(Symbol.Close)):
+							base_expression = FunctionCall(base_expression, [])
+						else:
+							arguments: List[Expression] = [parse_expression(lexer)]
+							while lexer.take_token(SymbolToken(Symbol.Comma)):
+								arguments.append(parse_expression(lexer))
+							lexer.assert_token(SymbolToken(Symbol.Close))
+							base_expression = FunctionCall(base_expression, arguments)
+					case Symbol.OpenSquare:
+						index = parse_expression(lexer)
+						lexer.assert_token(SymbolToken(Symbol.CloseSquare))
+						base_expression = ArrayAccess(base_expression, index)
 					case _: assert False, ("Unimplemented", symbol)
 			case _: assert False, ("Unimplemented", token)
 
@@ -184,7 +185,6 @@ def parse_unary(lexer: Lexer) -> Expression:
 	if lexer.take_token(SymbolToken(Symbol.Exclamation)): return Not(parse_unary(lexer))
 	if lexer.take_token(SymbolToken(Symbol.Plus)): return UnaryPlus(parse_unary(lexer))
 	if lexer.take_token(SymbolToken(Symbol.Dash)): return UnaryMinus(parse_unary(lexer))
-	if lexer.take_token(KeywordToken(Keyword.Return)): return Return(parse_unary(lexer))
 	return parse_primary(lexer)
 
 Precedence = Dict[Symbol, Callable[[Expression, Expression], Expression]]
@@ -220,7 +220,6 @@ def parse_ternary(lexer: Lexer) -> Expression:
 	return Ternary(expression, if_true, if_false)
 
 def parse_expression(lexer: Lexer) -> Expression:
-	if lexer.take_token(KeywordToken(Keyword.Return)): return Return(parse_expression(lexer))
 	if lexer.take_token(KeywordToken(Keyword.Lambda)):
 		parameters: List[TypeDeclaration]
 		if lexer.take_token(SymbolToken(Symbol.EqualArrow)):
@@ -301,6 +300,10 @@ def parse_statement(lexer: Lexer) -> Statement:
 	elif lexer.take_token(KeywordToken(Keyword.Continue)):
 		lexer.assert_token(SymbolToken(Symbol.Semicolon))
 		return ContinueStatement()
+	elif lexer.take_token(KeywordToken(Keyword.Return)):
+		expression = parse_expression(lexer)
+		lexer.assert_token(SymbolToken(Symbol.Semicolon))
+		return ReturnStatement(expression)
 	elif lexer.take_token(KeywordToken(Keyword.Do)):
 		body = parse_statement(lexer)
 		condition: Optional[Expression] = None
@@ -310,7 +313,6 @@ def parse_statement(lexer: Lexer) -> Statement:
 		return DoWhileStatement(body, condition)
 	elif lexer.take_token(KeywordToken(Keyword.Match)):
 		value = parse_expression(lexer)
-		lexer.assert_token(KeywordToken(Keyword.In)) # to prevent it from parsing it as a struct instantiation
 		lexer.assert_token(SymbolToken(Symbol.OpenCurly))
 
 		cases: List[Tuple[Expression, Statement]] = []
@@ -329,7 +331,7 @@ def parse_statement(lexer: Lexer) -> Statement:
 		body = parse_statement(lexer)
 		return ForLoop(variable, expression, body)
 	elif lexer.take_token(KeywordToken(Keyword.Import)):
-		file = parse_expression(lexer)
+		file = parse_string(lexer)
 		lexer.assert_token(SymbolToken(Symbol.Semicolon))
 		return Import(file)
 	elif lexer.take_token(KeywordToken(Keyword.Type)):
@@ -341,8 +343,10 @@ def parse_statement(lexer: Lexer) -> Statement:
 	elif lexer.take_token(KeywordToken(Keyword.Defer)):
 		statement = parse_statement(lexer)
 		return DeferStatement(statement)
-	elif lexer.check_tokenkind(KeywordToken) and not lexer.check_tokens(KeywordToken(Keyword.Return), KeywordToken(Keyword.Lambda)):
-		assert False, ("Unimplemented", lexer.next_token(), lexer.next_token(), lexer.next_token())
+	elif lexer.check_tokenkind(KeywordToken) and not lexer.check_token(KeywordToken(Keyword.Lambda)): # TODO: Maybe use '\' for lambda instead of a keyword
+		token = lexer.next_token()
+		assert isinstance(token.contents, KeywordToken)
+		raise SyntaxError(f"{token.loc}: Unexpected keyword: '{token.contents.keyword}'")
 	elif lexer.take_token(SymbolToken(Symbol.OpenCurly)):
 		statements: List[Statement] = []
 		while not lexer.take_token(SymbolToken(Symbol.CloseCurly)):

ppp_stdlib.py

@@ -2,7 +2,7 @@ from typing import Callable, Dict, List, Tuple
 
 from ppp_ast import Statements
 from ppp_object import Bool, EnumValue, Int, Object, Function, Str, TypeObject, Void, List as ListObject
-from ppp_types import Bool as BoolType, FunctionType, GenericType, Int as IntType, Str as StrType, Type, TypeType, VariableType, Void as VoidType, Object as ObjectType, UnionType, ListType
+from ppp_types import Bool as BoolType, FunctionType, GenericType, Int as IntType, Str as StrType, Type, TypeType, VariableType, Void as VoidType, Object as ObjectType, ListType
 
 
 def PythonFunction(name: str, parameters: List[Tuple[str, Type]], return_type: Type, func: Callable[..., Object]) -> Object:
@@ -41,7 +41,14 @@ def len_impl(list_: Object) -> Object:
 		case _: assert False, ("Unimplemented", list_)
 	assert False
 
-Len = PythonFunction("len", [('list', UnionType([ListType(VariableType("")), StrType]))], IntType, len_impl)
+# TODO: Use polymorphism to make this work for both list<T> and str
+Len = PythonFunction("len", [('list', ListType(VariableType("")))], IntType, len_impl)
+
+def str_len_impl(str_: Object) -> Object:
+	assert isinstance(str_, Str)
+	return Int(len(str_.str))
+
+StrLen = PythonFunction("strlen", [('string', StrType)], IntType, str_len_impl)
 
 def str_to_int_impl(str_: Object) -> Object:
 	assert isinstance(str_, Str)
@@ -100,6 +107,7 @@ variables: Dict[str, Object] = {
 	'debug_print': DebugPrint,
 	'read': Read,
 	'len': Len,
+	'str_len': StrLen,
 	'str_to_int': StrToInt,
 	'none': NoneObj,
 	'range': Range,

ppp_tokens.py

@@ -1,7 +1,5 @@
 from dataclasses import dataclass
 from enum import Enum
-from typing import List, Literal, Tuple, Union
-
 
 class Keyword(Enum):
 	Enum = 'enum'
@@ -25,6 +23,8 @@ class Keyword(Enum):
 	Type = 'type'
 	Defer = 'defer'
 
+	def __str__(self) -> str: return self._value_
+
 class Symbol(Enum):
 	Open = '('
 	Close = ')'
@@ -62,6 +62,8 @@ class Symbol(Enum):
 	Tilde = '~'
 	Carot = '^'
 
+	def __str__(self) -> str: return self._value_
+
 @dataclass
 class KeywordToken:
 	keyword: Keyword
@@ -88,18 +90,19 @@ class SymbolToken:
 @dataclass
 class EofToken: pass
 
-TokenContents = Union[
-	KeywordToken,
-	IdentifierToken,
-	NumberToken,
-	StringToken,
-	SymbolToken,
-	EofToken
-]
+TokenContents = KeywordToken | IdentifierToken | NumberToken | StringToken | SymbolToken | EofToken
+
+@dataclass
+class Location:
+	file: str
+	line: int
+	col: int
+
+	def __repr__(self) -> str:
+		return f"{self.file}:{self.line}:{self.col+1}"
 
 @dataclass
 class Token:
-	line: int
-	col: int
+	loc: Location
 	value: str
 	contents: TokenContents

ppp_types.py

@@ -1,7 +1,7 @@
 
 from abc import ABC, abstractmethod
 from dataclasses import dataclass
-from typing import Dict, List, Tuple, Union
+from typing import Dict, List, Tuple
 
 import sys
 sys.setrecursionlimit(1000)
@@ -40,10 +40,6 @@ class Type(ABC):
 			case VariableType(self_name), VariableType(other_name):
 				return self_name == other_name
 			case _, VariableType(""): return True
-			case type, UnionType(types):
-				for union_type in types:
-					if type.is_subtype_of(union_type): return True
-				return False
 			case BoolType(), BoolType(): return True
 			case type, ObjectType(): return True
 			case type_a, type_b if type_a.__class__ != type_b.__class__: return False
@@ -164,41 +160,10 @@ class FunctionType(Type):
 		is_new_return_type, new_return_type = self.return_type.new_fill(types, stack+[id(self)])
 		return (is_new_arguments or is_new_return_type, FunctionType(new_arguments, new_return_type))
 
-@dataclass
-class UnionType(Type):
-	types: List[Type]
-
-	def fill(self, types: Dict[str, Type], stack: List[int]) -> Type:
-		if id(self) in stack: return self
-		self.types = [type.fill(types, stack+[id(self)]) for type in self.types]
-		return self
-
-	def new_fill(self, types: Dict[str, Type], stack: List[int]) -> Tuple[bool, Type]:
-		is_new, new_types = self.new_fill_list(self.types, types, stack)
-		return (is_new, UnionType(new_types))
-
-	def represent(self) -> str: return '('+'|'.join([type.represent() for type in self.types])+')'
-
 class ObjectType(Primitive):
 	def represent(self) -> str: return 'object'
 Object = ObjectType()
 
-@dataclass
-class ReturnType(Type):
-	type: Type
-
-	def represent(self) -> str: return f"return<{self.type.represent()}>"
-
-	def fill(self, types: Dict[str, Type], stack: List[int]) -> Type:
-		if id(self) in stack: return self
-		self.type = self.type.fill(types, stack+[id(self)])
-		return self
-
-	def new_fill(self, types: Dict[str, Type], stack: List[int]) -> Tuple[bool, Type]:
-		assert id(self) not in stack
-		is_new, new_type = self.type.new_fill(types, stack+[id(self)])
-		return (is_new, ReturnType(new_type))
-
 
 num_expressions: int = 0
 @dataclass